Across adult online training and small‑group workplace sessions that both use an attempt–worked‑example cycle, does briefly replaying a learner’s own prior unguided attempt (e.g., showing their earlier draft or transcript before the new attempt) improve long‑term retention, transfer of skills, and calibration more than simply spacing practice without replay, or does this replay mainly increase cognitive load without additional learning benefits?

Answer

Briefly replaying a learner’s own prior unguided attempt just before a new attempt is plausibly a small net positive for long‑term retention, transfer, and calibration compared with spacing alone, provided the replay is short, framed as a comparison tool, and not overloaded with commentary. In that constrained form, it:

• likely yields small gains in long‑term retention and transfer (by cueing retrieval and highlighting changes in strategy), and
• moderately improves calibration (by making discrepancies between felt and actual performance salient),
• while adding only modest extra cognitive load.

However, if replay is long, detailed, or evaluative (e.g., full transcripts with heavy markup, or multiple old attempts at once), it tends to increase extraneous cognitive load and distraction with little added benefit over simple spacing. In low‑prior‑knowledge learners or low‑safety small‑group settings, such heavy replay can even impair performance and engagement.

Design implications:

• Use very short, focused self‑replay (e.g., a snapshot of key steps or errors) as an optional or default pre‑cue before new attempts.
• Pair it with one or two simple prompts (e.g., “What would you change this time?”) rather than extensive analysis.
• Avoid replay formats that require substantial rereading or self‑critique in real time, especially for novices or anxious group members.

Overall, replay is best treated as a lightweight metacognitive cue layered on top of spacing, not as a heavy review phase; in that form, its benefits modestly outweigh its cognitive‑load costs for most learners.